Embodied AI Glossary中文Scale AI
AdvancedA US AI data-labeling company that supplies training data for large models and autonomous driving.
Scale AI was founded in San Francisco in 2016 by Alexandr Wang and Lucy Guo. Its core business is data annotation and model evaluation: it organizes large teams of human labelers to prepare training data for autonomous driving, computer vision, and large language models, including the human-preference data needed for reinforcement learning from human feedback (RLHF). In June 2025, Meta paid over $14 billion for a 49% non-voting stake, and Wang joined Meta to lead its superintelligence team, with Jason Droege taking over as CEO; in March 2026 the company launched a research division called Scale Labs. In embodied AI, Scale represents the “data-service provider” role, and it is reported to be expanding into robotics and physical-AI data as well.
ExampleScale's early customers were mostly autonomous-driving companies, which used it to label vehicles and pedestrians in LiDAR point clouds and camera images.
- Related
- Data Annotation · Embodied AI Data Service Provider · Reinforcement Learning from Human Feedback · Autonomous Driving · OpenAI
- Sources
- Scale AI - Wikipedia
- As of
- 2026-03